Dashboards (see Note 1) measure performance. Strategy management solutions use scorecards (see Note 2) and strategy maps to go beyond dashboards and to correlate objectives with one another and with their underlying performance indicators, allowing performance to be managed. A strategy map seeks to establish cause and effect among factors that are key to financial success by linking strategy formulation to tactical execution.
1 These associations are typically discovered and maintained through casual observation, intuition and gut feel.
This method is challenging, because the relationships among these metrics are often complex, dynamic and time-intensive. If not continually analyzed and updated, a strategy map will eventually become meaningless. As a result, the relationships among metrics are typically created and maintained at either higher aggregated levels or lower operational levels. These approaches have value; however, a more comprehensive enterprise metrics framework (see Note 3) will more accurately link strategic goals and operational activities.
Some organizations have worked to address this strategy-mapping problem by applying advanced statistical and information visualization techniques to gain additional insights into the intricate cause-and-effect relationships among measures. This can be advantageous, because some of these relationships aren't intuitive enough to detect or explain without applying mathematical rigor, and are too complex to visually represent using simple hierarchical structures.
Traditionally, advanced statistical and information visualization techniques have been out of scope for most scorecarding endeavors. Advanced solutions for strategy management embed statistical and sophisticated information visualization capabilities to help end users identify, explain and maintain performance measure relationships. These capabilities promise both complex causal analysis and simultaneous scorecard transparency; these features are often mutually exclusive.
In recognition of this opportunity for improvement, CPM vendors — such as IBM (CFO Performance Dashboard v.3 Advanced Edition), a services-led IBM offering (Global Business Services) and SAS (SAS Strategy Management) — embed certain types of advanced analytics (see Note 4) into their products. However, what must organizations do to realize benefits from them? Strategy mapping requires a multifaceted effort of people, processes and technology; however, to leverage advanced analytics to improve strategy mapping, organizations must understand how it can help. Specifically, users need to know
- How these methods can extend their organization's strategy management ability
- What is the likelihood this functionality can leverage your existing data
- When to apply advanced analytics to strategy mapping
- Why it's necessary to foster a culture of data-driven decision making
- Where to start
Advanced Analytics Can Extend Strategy Management Capabilities
Advanced analytics can assist with three major strategy-mapping challenges:
- Identifying performance measurement relationships
- Explaining the nature of their relationships to foster consensus
- Helping to maintain them
This maintenance involves ensuring that these metrics don't proliferate (only metrics with a direct lineage to key performance drivers should be perpetuated), as well as the ongoing monitoring of their relationships.
Identifying and explaining the correlations among metrics is a challenging task. Gartner estimates that 80% of the effort behind a scorecard or dashboard initiative consists of defining the metrics and finding the right data (see
"Just Give Me a CPM Dashboard" ). Most strategy management products require a predetermined knowledge of all performance measurements and an understanding of the relationships between them and their attributes (whether they're lagging or leading indicators, their weighting, etc.). Embedded functionality that helps identify these associations, represent them visually and provide a fact-based foundation for meaningful discussion could improve the accuracy and transparency of strategy maps.
Once established, maintaining the links between performance measurements is difficult. A major point of strategic scorecard failure can be their maintenance, especially during instances of sponsorship, business or environmental change. Modifications to executive management, merger and acquisition (M&A) activity, and disruptive economic, technical or competitive alterations are just a few examples of the events types that can result in changes to relationships among performance measures. Traditional strategy management products largely assume consistency over time. The introduction of analytics to monitor the association of these measures can help streamline this maintenance task.
What is the likelihood this functionality can leverage your existing data? Identifying correlations among performance measurements may be useful, but, just as often, a simple correlation analysis may result in false positives. For example, a regression analysis used to quantify the relationships among two or more metrics may indicate a strong correlation, even though there is no causal relationship. Such an example may result in nonsensical conclusions, such as on-time delivery performance having a strong positive correlation with currency fluctuations.
Analyzing the relationships among performance measurements requires statistical methods that go beyond correlation analysis to identify the causal relationships among these measurements and information visualization functionality to graphically explain them. To do this, these embedded analytic techniques also need to leverage meaning from data changes over time; so, in general, the more historical data available, the better.
For these calculations to be effective, data also needs to have an appropriate level of accuracy and granularity, so that the components that make up the measurement can be analyzed. For strategy management, the level of history, quality and granularity of the data used is largely determined by the integration capabilities of your CPM products and/or the design of your CPM applications. Additional regulatory requirements, new capabilities that allow the integration of strategic and operational planning, integrated tax provisioning needs, and requests for more-meaningful internal and external reporting have expanded the data landscape of CPM systems.
CIOs and business managers need to work together to resolve the data issues underpinning meaningful metrics. In addition to addressing data shortcomings, you need to understand that certain business data characteristics can also affect statistical interpretations. Data with these characteristics will require additional manipulation and harmonization. The common temporal characteristics of business data that may need to be addressed include the following.
Restated Data/Data Consistency
Financial results often need to be restated. For example, error corrections, regulatory requirements or M&A activity can affect performance-related statistical interpretations. In addition, data from businesses that are no longer reported on or are reported on differently over time may skew statistical analyses of performance measure linkages. Another consistency issue can result from changes in account definition, changes to accounting treatments and policies, or incorrect or inconsistent use of accounts or other metadata. These conditions are likely to require management intervention to make appropriate interpretations of statistical results, or possible data modifications and additions to aid statistical testing.
Data Affected by Specific Environmental Events
The relationship among measurements can be distorted by specific events. For example, an unprecedented natural disaster or economic event may cause performance to artificially improve or decline, depending on the type of business or other circumstances (e.g., Eurozone interest rate and currency fluctuations). The effects of these events may cause periods of uncharacteristic activity.
Human decision making is subject to its own set of biases, especially in the face of unique fortuitous or disastrous change. In these instances, a dispassionate, data-driven approach can act as a bellwether to augment the decision-making process and identify these anomalies. As always, management needs to understand the capabilities of the statistical methods used and to properly interpret their results.
Data Volatility
Data with extensive variation over time is generally less useful than data containing stable patterns that can be used for interpretation. This is particularly problematic when shorter time periods are measured, or when historical data is offloaded to less accessible archives. Business units operating in unstable markets, competitive situations and those experiencing other erratic factors in the business environment can cause this type of volatility. Such variations may be acute for M&A data, which typically contains only recent performance or newly acquired assets. There are also instances when advanced analytics cannot be applied. For example, traditional methods would need to be relied on for new, unique operations in which there is no historic data for statistical interpretation.
When to Apply Advanced Analytics to Strategy Mapping
Although applicable to many performance measure association analysis efforts, this approach will be most effective when the strategy-mapping endeavor is complex, and when data and culture can support their use. For example, Gartner recommends the use of an enterprise metrics framework to link overall strategic goals with operational activities (see Figure 1). This provides a common set of metrics that can be consistently measured and managed across an organization and links the achievement of corporate goals and objectives with operational activities (see
"Tutorial for Creating an Enterprise Metrics Framework" [Note: This document has been archived; some of its content may not reflect current conditions]).
Figure 1. An Enterprise Metric Framework
Source: Gartner (February 2012)
Because this framework includes performance metrics across the enterprise, it may lead to a level of complexity that is difficult to implement and maintain using a traditional observational approach. This approach can benefit from advanced analytics and a data-driven decision-making culture, because the more extensive and complex the correlations are, the greater the risk that the resulting scorecard will lose transparency and management understanding, doing more harm than good.
Advanced analytics can help identify, explain and maintain these relationships. More importantly, it can help provide new insights. A great deal of the value of a strategy-mapping exercise is in related discussions and the resultant insights gained. Even an unsuccessful correlation analysis can result in new management understanding.
A Culture of Data-Driven Decision Making Should Be Supported
CEOs regard data-driven decision-making capabilities as having the most potential strategic value to the business (see Figure 2 and
"Executive Advisory: CEO and Senior Executive Survey" ).
Figure 2. Expected Value of Various Technology-Enabled Capabilities to Respondent Organizations, 2011-2014
Source: Gartner (February 2012)
